This invention relates to precursors and methods for making thin films that are useful for the fabrication of integrated circuits (“IC”). In particular, this invention relates to thin films that are created by polymerizing fluorinated ethylinic precursors with fluorinated benzocyclobutane precursors, fluorinated biphenyl or fluorinated dieneone precursors. The resultant thin films have increased compositional strength, a low-dielectric constant (“∈”), and are stable at high temperatures.
As integrated circuits (“ICs”) have become progressively more microminiaturized to provide higher computing speeds, current dielectric materials used in the manufacturing of the ICs have proven to be inadequate in several ways. These materials, for instance, have high dielectric constants, difficulty to use in the manufacturing process, have inadequate thermal instability and generate of toxic by-products. ICs are made by depositing layers of elements and/or compounds on a semiconductor wafer using a variety of techniques that are well known in the art of fabricating such devices. Specialized material are used to isolate layers on the IC and reduce the charge (i.e. capacitance) that can be stored in between conducting elements of the IC. To reduce the potential capacitance in certain layers, it is preferable that the materials have a low dielectric constant (“∈”). Low dielectric constant materials can be deposited by a variety of methods, including spin-on and chemical vapor deposition (CVD). The composition and characteristics of the dielectric materials are determined from its precursors as well as the processes and reactions such precursors undergo while being integrated into the IC. As used herein, spin-on refers to the IC manufacturing process whereby the substrate is rotated about an axis perpendicular to its surface while, or immediately after, a coating material is applied to the surface. As ICs become smaller and more functional, a dielectric material with ∈ that is 2.7 or lower will be required.
Other properties such as thermal stability, compositional integrity and process compatibility are important factors that must be considered when integrating a dielectric material into an IC. For example, a dielectric material should retain its integrity during the processes involved in IC fabrication. These processes include reactive ion etching (“RIE”) or plasma patterning, wet chemical cleaning of photoresist, physical vapor depositions (“PVD”) of barrier materials and cap layers, electroplating and annealing of copper (“Cu”) and chemical-mechanical polishing (“CMP”) of copper. In addition, the dielectric should have sufficient dimensional stability. Interfacial stresses resulting from a coefficient of thermal expansion (“CTE”) mismatch between the dielectric and barrier material should not induce structural failure of the barrier material during and after annealing of copper. In addition, the interfacial adhesion of dielectric and the other barrier material should be sufficient to overcome interfacial and shear stresses and warrant good adhesion after annealing and CMP of copper. Corrosive organic elements used for IC processing can cause interfacial corrosion of the barrier material, and it is essential that the dielectric material does not allow the organic elements to diffuse into the barrier material layer. In addition, to maintain its electrical integrity after fabrication of the ICs, the dielectric should be free from contamination by the barrier material. Furthermore, the interfaces of dielectric and the barrier material should be free from moisture and no ionic migration occurs when the ICs are operating under electrical bias.
Dielectric materials that have been traditionally used in ICs were either solid or porous thin films. There are advantages and disadvantages to each. For example, the advantages of solid dielectric materials include: higher dimensional and structural integrity and better mechanical strength than porous dielectric materials, but the disadvantage is higher dielectric constant. In contrast, the advantage of porous dielectric materials is lower dielectric constant due to the presence of air inside tiny pores of these materials. Current solid materials are unable to achieve stability, integrity and strength with a dielectric constant below 2.7.
The “solid” polymer films or “pin-hole free” films contain voids that can generally range between 3 to 5 volume % of the films. However, the average void sizes in a cross-section of a well prepared “pin-hole free” or “solid” films are only few Angstroms. It is critical that the pore sizes of the thin films be relatively small in order to be useful for fabrication of current or future generation of ICs. For example, the pore sizes should be less than the mean free path (i.e. 50 to 100 Angstroms) of the barrier material, which is typically Tantalum (“Ta”).
The removal of solvents or sacrificing materials can result in additional porosity and low dielectric constant in “pin-hole free” polymer films. However, when the sacrificing materials have different compatibilities with the polymer matrix, the result can lead to polymer aggregation and pore sizes larger than 100 Angstroms. The resulting thin film dielectric has poor mechanical properties due to localized degradation caused by large pores or their aggregates. The presence of pores in theses dielectric materials normally results in holes oh newly formed surfaces, thus making subsequent depositions of a continuous, thin <50–100 Å) barrier layers and copper seed layers very difficult if not impossible. Additional problems with traditional porous thin films are they often exhibit reliability problems due to the inclusion of barrier metal inside the dielectric layer, as occurs after PVD of Ta. Porous dielectric materials are also difficult to integrate into IC fabrications that involve a CMP process. To further complicate the process, large surface areas in porous films lead to high water adsorption that can limit the electrical reliability of the IC.
Precursors such as Bicyclobutene (“BCB”) can be used to make thin films in a copper dual damascene structure without the need for a barrier layer such as Ta, however, the dielectric constant of BCB is greater than 2.7. Introduction of air bubbles into the BCB during the process can increases porosity and a consequential decrease of the dielectric constant. At 20% porosity, BCB has a dielectric constant of about 2.3. Unfortunately, the porous BCB and other dielectric materials that can achieve a ∈≦2.4 are too soft for CMP and not suitable for fabrication of current and future ICs.
Plasma polymerization of fluorinated precursor molecules has also been described. For example, Kudo et al., Proc. 3d Int. DUMIC Conference, 85–92 (1997) disclosed polymers made from C4F8 and C2H2 with a dielectric constant of 2.4. The polymers had a glass transition temperature (“Tg”) of 450° C. However, despite its low leakage current due to presence of sp3C—F bonds, a low thermal stability occurred due to presence of sp3C—F and sp3C-sp3-C bonds in the films. Thus, these fluorinated polymers are unable to withstand the prolonged high temperatures necessary for IC manufacture. In addition, LaBelle et al, Proc, 3d Int. DUMIC Conference, 98–105 (1997) also described the use of CF3—CF(O)—CF2 precursors in a pulsed plasma CVD process, which resulted in some polymer films with a dielectric constant of 1.95. However, in spite of the low dielectric constant, these polymer films also had a low thermal stability due to presence of sp3C-sp3C and sp3C—F bonds in these films.
Other fluorinated compounds described by Wary et al, (Semiconductor International, June 1996, 211–216) used the dimer precursor, (α, α, α1, α1), tetrafluoro-di-p-xylylene (i.e. {—CF2—C6H4—CF2—}2) and a thermal CVD process to manufacture Parylene AF-4™, which has the structural formula: {—CF2—C6H4—CF2—}n. Films made from Parylene AF44™ have a dielectric constant of 2.28 and have increased thermal stability compared to the above-mentioned dielectric materials. Films made of Parylene AF-4™ lost only 0.8% of its weight over a 3-hour period at 450° C. under a nitrogen atmosphere. However, there are disadvantages to the known methods the manufacture of the fluorinated poly (para-xylylenes), or Parylene AF44™. First, the manufacture of their precursors is inefficient because the chemical reactions have low yields, and the process is expensive and produces toxic byproducts. Further, it is difficult to eliminate redimerization of the reactive intermediates. When deposited along with polymers, these dimers decrease the thermal stability and mechanical strength of the film.
In our co-pending applications, we have disclosed some pin-hole-free polymer dielectric that can be prepared from transport polymerization process. These dielectric materials consist of sp2C—F and hyperconjugated sp3C—F in their polymer chains, thus they have ∈≦2.4, and they are thermally stable for fabrication of future ICs. Herein, we describe precursors and processes for making thin films from precursors that results in polymers with low dielectric constant, improved compositional strength and high temperature stability that should provide low cost alternatives for fabrication of miniaturized ICs.